Active tools for parameter adapting and failure monitoring of integrated data-processing systems

This paper examines the problem of parametric identification of systems that use the Kalman data model. In practical terms, it focuses on the task of estimating errors of the inertial navigation system (INS). On the whole, the INS error model contains many parameters that are either unknown beforehand or can change abruptly. To demonstrate a well-argued solution for the tasks at hand and to break it down step-by-step, the paper uses a specific streamlined example of such a model. The solution lies in constructing a measurable Auxiliary Performance Index (API), whose minimum in adjustable parameters ensures that the adaptive filter matches the optimal Kalman filter, given that (theoretically) the estimates are unbiased.